In the first case, his argument is that the interest rate (a price, administered to be sure) would normally equilibrate savings and investment at something like full employment. It’s the barrier to that price’s adjustment in the form of the zero lower bound that causes income to adjust instead, leading to the Great Recession (and the need for fiscal policy). Similarly trade flows are normally equilibrated by the adjustment of the exchange rate, a price. It’s barriers to that price’s adjustment, in the form of the Euro and the RMB-dollar peg, that cause income to adjust instead, leading to austerity in peripheral Europe and unemployment in the United States.
From a Post Keynesian perspective, these are kludges that get good conclusions from bad premises.
From a Post Keynesian (or for that matter Keynesian straight-up) perspective, flexible interest rates and exchange rates have never reliably delivered macroeconomic balance. Income adjustments aren’t a once-in-a-lifetime feature of the current conjuncture, they’re a routine and central feature of capitalism.
New Keynesian economists like Krugman can see that macroeconomic reality today doesn’t conform to the textbook, where prices smoothly converging to market-clearing levels. (Well, maybe to the 1978 edition.) But they’re not going to throw the textbook away, so the departures are explained as a series of ad hoc special cases. And so the textbook has a way of sneaking back in whenever their attention is elsewhere. That’s why Krugman insists something big changed when the federal funds rate hit zero, even though the federal funds rate has been more or less disconnected from most longer rates for a decade or more. And that’s why he insists that the Asian crisis countries were better off than Greece, etc. because they could devalue their currencies instead of resorting to austerity, when it seems clear that devaluations contributed little to Asian countries’ improved current account balances after 1997; they drastically cut domestic spending, just as peripheral Europe is being forced to. When he’s looking right at a non-price adjustment mechanism, he can see it; but wherever he’s not looking, he assumes that prices are doing their thing.
Or at least that’s how it looks to me.